If "acpi_osi=Linux" works better, pleas e notify linux-acpi@vger.kernel.org

If "acpi_osi=Linux" works better, pleas e notify linux-acpi@vger.kernel.org

[Dennis Jansen]

Hi,

it works a lot better with acpi_osi=Linux in 2.6.24 x64. Wireless is
much more stable and the power consumption decreases by about 2 W.

I already sent you my dmidecode and acpidumps in the other mail a few days ago.

Cheers,
Dennis

Re: If "acpi_osi=Linux" works better, pleas e notify linux-acpi@vger.kernel.org

[Len Brown]

On Sunday 10 February 2008 09:37, Dennis Jansen wrote:
> Hi,
> 
> it works a lot better with acpi_osi=Linux in 2.6.24 x64. Wireless is
> much more stable and the power consumption decreases by about 2 W.

Are you sure the only difference is booting with "acpi_osi=Linux"?

In both BIOS V1.21 and V1.31 it appears to be a no-op.
The only way I could imagine it could have an effect is if
somethign in SMM is looking at the LINX variable.

Based on past lack of Linux awareness from this particular vendor,
the chances of that happening seem quite small.

thanks,
-Len

Re: If "acpi_osi=Linux" works better, pleas e notify linux-acpi@vger.kernel.org

[Dennis Jansen]

On Feb 11, 2008 3:29 AM, Len Brown <lenb@kernel.org> wrote:
> In both BIOS V1.21 and V1.31 it appears to be a no-op.
> The only way I could imagine it could have an effect is if
> somethign in SMM is looking at the LINX variable.
>
> Based on past lack of Linux awareness from this particular vendor,
> the chances of that happening seem quite small.
>
> thanks,
> -Len
>
It was the suspend/resume process and the subsequent use of the C2
state that made the difference.
And the wireless stability is also another issue.
Sorry, guess it was wishful thinking.